SUMMARY
The nervous system spread of alpha-synuclein fibrils is thought to cause Parkinson’s disease (PD) and other synucleinopathies, yet the mechanisms underlying internalization and cellular spread are enigmatic. Here we use confocal and super-resolution microscopy, subcellular fractionation and electron microscopy (EM) of immunogold labelled alpha-synuclein preformed fibrils (PFF) to demonstrate that this fibril form of alpha-synuclein undergoes rapid internalization and is targeted directly to lysosomes in as little as 2 minutes. Uptake of PFF is disrupted by macropinocytic inhibitors and circumvents classical endosomal pathways. Immunogold-labelled PFF are seen at the highly curved inward edge of membrane ruffles, in newly formed macropinosomes, in multivesicular bodies and in lysosomes. While most fibrils remain in lysosomes, a portion is transferred to neighboring naïve cells along with markers of exosomes. These data indicate that PFF use a unique internalization mechanism as a component of cell-to-cell propagation.
INTRODUCTION
A classic hallmark of Parkinson’s disease (PD) is the formation of Lewy Bodies (LBs). First discovered in 1912 (Lewy, 1912), LBs are cytoplasmic inclusions composed of fragments of membranous organelles and filamentous proteins (Shahmoradian et al., 2019; Tanaka et al., 2004). LBs are also observed in PD-related disorders such as Lewy body dementia. Alpha-synuclein (α-syn), encoded by the SNCA gene, is a major component of LBs and is implicated in their formation (Conway et al., 1998; Spillantini et al., 1997). α-syn appears to function in membrane trafficking and membrane curvature (Fortin et al., 2004; Jao et al., 2004; Nakamura et al., 2008; Vargas et al., 2017), and increased knowledge of its protein structure and the conformation of altered variants has led to enhanced understanding of α-syn misfolding and aggregation in disease (Lee et al., 2002; Li et al., 2002; Sandal et al., 2008).
Since the early 2000s, the Braak hypothesis has stated that PD develops with the spread of α-syn through the brain (Braak et al., 2003). This model gained traction with the observation that proteinaceous inclusions spread from brain tissue into implanted embryonic stem cells (Li et al., 2008), a result subsequently confirmed in animal models (Desplats et al., 2009b; Li et al., 2010; Recasens and Dehay, 2014; Recasens et al., 2018). Propagating pathology is also observed following injection of the fibril form of α-syn into localized brain regions of mice (Betemps et al., 2014; Luk et al., 2012; Masuda-Suzukake et al., 2014). This tendency to spread, along with the ability of α-syn fibrils to disrupt the conformation of endogenous α-syn, has earned the protein a label as a prion-like (Masuda-Suzukake et al., 2013; Mougenot et al., 2012). However, questions remain regarding the cell biological itinerary of α-syn propagation, notably the mode of cellular entry (Fenyi et al., 2019; Gelpi et al., 2014; Nakamura et al., 2015; Uemura et al., 2018; Yan et al., 2018).
Different mutations in the SNCA locus have varying penetrance that may correlate to the propensity of α-syn to aggregate (Lazaro et al., 2016; Rutherford et al., 2014). α-syn concentration is also a factor in aggregation as increased protein levels enhance aggregation, be it mutant or wild-type protein (Fink, 2006; Manning-Bog et al., 2002; Uversky, 2007). Further, α-syn preformed fibrils (PFF) are more effective at spreading and seeding α-syn aggregates than monomeric and other fibrillar forms (Conway et al., 2000; Lam et al., 2016). PFF comprise a heterogeneous number of α-syn monomers with various structural conformations (Pieri et al., 2016). Early studies using PFF revealed their ability to seed and induce PD pathology in cultured cells (Luk et al., 2009); hence, developing a protocol for consistent production of PFF was a significant contribution (Volpicelli-Daley et al., 2014; Volpicelli-Daley et al., 2011).
A key question in PD research relates to the means by which α-syn fibrils enter cells (Bieri et al., 2018). Several studies have concluded that α-syn endocytosis is clathrin-dependent, based on the use of dynamin and clathrin inhibitors, dynasore and pitstop, respectively, and clathrin heavy chain (CHC) colocalization (Konno et al., 2012; Liu et al., 2007; Rodriguez et al., 2018). It is generally thought that PFFs are then trafficked to lysosomes via the endosomal system over the course of tens of minutes to multiple hours. We found approximately 40 studies examining α-syn fibril endocytosis with variable conclusions (Table S1). Some of the variability may arise from examining α-syn internalization at longer time courses and not immediately after its addition to cells. Moreover, dynamin inhibition is not synonymous with inhibition of clathrin-mediated uptake and the specificity of dynasore and pitstop as inhibitors of clathrin-mediated endocytosis (CME) has come under scrutiny as both have off-target effects (Gu et al., 2010; Oh et al., 1998; Park et al., 2013; Pelkmans et al., 2002; Preta et al., 2015). Apart from CME, it is suggested that α-syn enter cells via direct permeation of the plasma membrane (Danzer et al., 2007), via the formation of tunneling nanotubes that allow direct connections between cells (Dieriks et al., 2017), or through caveolae-dependent endocytosis (Madeira et al., 2011). Thus, there appears to be no consensus nor definitive evidence regarding the nature of α-syn endocytosis.
We thus sought to examine endocytosis of PFF immediately after their addition to cells. Surprisingly, PFF are internalized rapidly and appear in lysosomes within 2 min, bypassing conventional endosomal trafficking pathways. We confirmed this result in multiple cell lines, primary human cells, and neurons derived from induced pluripotent stem cells (iPSC). The internalization is not dependent on clathrin but instead uses macropinocytosis. We used gold-labeled PFF and EM and discovered PFF in membrane ruffles that form macropinosomes and in lysosomes. We also detect gold-labeled PFF on the outer edges of inward invaginating vesicles in endosomes and on the outside of vesicles within multivesicular bodies (MVB). While a portion of PFF remain in lysosomes for a long period, a smaller portion are transferred to naïve cells along with markers of MVB. Thus, our data unveils a unique form of macropinocytosis that mediates internalization of PFF and allows for endocytosis to be coupled to release.
RESULTS
PFF are rapidly endocytosed to lysosomes
We used fluorescently labeled PFF (Del Cid Pellitero et al., 2019; Maneca et al., 2019) (Fig. S1A/B) with live-cell imaging and a trypan blue exclusion assay (Karpowicz et al., 2017) that quenches extracellular PFF fluorescence (Fig. S1D) to examine PFF internalization at the earliest possible time points. PFF are rapidly internalized in U2OS cells and are targeted to lysosomes where they colocalize with Lysosomal Cytopainter within as little as 2 min (Fig. 1A). Similar results are seen when incubating cells with PFF at 4°C for 30 min and then transferring the cells to 37°C (Fig. S1E). We confirmed the live imaging findings in fixed samples of U2OS cells. We confirmed the live imaging findings in fixed samples of U2OS cells. PFF are significantly internalized within 2 min and continue to accumulate within cells for up to 60 min (Fig. 1B & S1H). Nearly all labeled PFF that enter cells are colocalized with LAMP1-TurboRFP by 2 min, and this stays stable with all PFF trafficking to lysosomes) for up to 60 min (Fig. 1C). As fixation makes cells permeable to trypan blue, for fixed cell experiments, we used trypsin to proteolyze off extracellular PFF (Fig. S1F/G). The rapid uptake of PFF and transport to lysosomes within 2 min was also observed in U87 (Fig. S2A/D) and U251 (Fig. S2B/E) glioblastoma cells. U2OS cells continue to accumulate PFF for up to 72 h (Fig. 1D).
To confirm the rapid transport of PFF to lysosomes, we incubated U2OS cells stably expressing HA-TMEM192-RFP, a lysosomal protein, with PFF for various time periods, then lysed the cells and purified lysosomes using HA-magnetic beads (Abu-Remaileh et al., 2017). The enrichment of lysosomes was confirmed in the immunoprecipitated fractions with antibodies recognizing LAMP1 and 2 (Fig. 1E). Rab7 was also enriched in the lysosome fractions, whereas Rab5 and LRRK2 were not detected. CD63/LAMP3, a marker of lysosomes and MVB, was the most highly enriched marker. PFF were enriched in the lysosome fractions as early as 2 min, confirming their rapid transport to lysosomes (Fig. 1E). The enriched lysosomes were placed on coverslips and visualized through the fluorescently-labelled PFF (Alexa Fluor 488) and the RFP tag on lysosomes (Fig. 1F). PFF were detected in the lumen of the lysosomes, which was most readily seen using STED super-resolution microscopy (Fig 1G). Fluorescently conjugated PFF of various sizes (Fig 1H) were administered to U2OS cells revealing no significant difference in uptake in samples ranging from 50-100 nm (Fig 1I). Thus, PFF are rapidly endocytosed and transported to lysosomes in as little as 2 min, an unprecedented time frame for lysosomal targeting.
Rapid transfer of PFF to lysosomes in nervous system cells including human dopaminergic neurons
Human cortical neurons derived from iPSCs (Fig. S1I) were incubated with fluorescently labeled PFF and colocalization with lysosomes was examined using a fixable form of Lysotracker. PFF were rapidly endocytosed in the neurons with lysosomal colocalization observed as early as 2 min (Fig. S2C/F/G), although lysosomes continued to accumulate PFF for up to 30 min. In both human dopaminergic neural progenitor cells (NPC) derived from iPSCs and in dopaminergic neurons derived from the NPC, PFF rapidly internalize and are detected at lysosomes within 2 min (Fig. S1I, 2A, B, D, E). Similar results are seen in human astrocytes (Fig. S1I, 2C/F). Thus, in multiple cell types, including those from the human nervous system, PFF are rapidly transported to lysosomes.
PFF were added to U343 glioblastoma cells for 24 h followed by a change to fresh media. PFF are trafficked to lysosomes and remain in the organelle, accumulating for up to 10 days (Fig. S2J). In contrast, internalized epidermal growth factor (EGF) dissipates from lysosomes (Fig. S2J). Internalized PFF remain in lysosomes in human dopaminergic neurons (Fig. 3A) and in human astrocytes (Fig. 3B), days after extracellular PFF has been removed from the cell media. The fluorescence intensity of PFF over 14 days was investigated in human astrocytes (Fig 3C). Quantification of the fluorescence showed a significant reduction in PFF fluorescence after day 1 (Fig 3D); however, a considerable amount of PFF remains in lysosomes at 14 days.
PFF are trafficked to lysosomes avoiding early/recycling endosomes
Considering the speed at which PFFs reach lysosomes it seems unlikely that PFF follow endosomal pathways to lysosomes (Lee et al., 2005; Lee et al., 2008; Lee et al., 2016) as endosomal maturation generally takes 10-15 min (Huotari and Helenius, 2011). Transferrin (Tf), a well-studied marker of early and recycling endosomes (Trischler et al., 1999) was added along with PFF and we observed no co-localization as 2 min and for as long as 30 min (Fig. S3A/D). Moreover, internalized PFF do not colocalize with EEA1, a marker of early endosomes (Mills et al., 1998), even at time points as early as 2 min (Fig. S3B/E). PFF colocalize with the late endosome/lysosome markers Rab9 and LAMP1 but show little co-localization with early and recycling Rabs, 4, 5 and 8 (Fig.S3C, F/G). Thus, PFF appear to reach lysosomes independently of the early and recycling endosomal systems.
Lysosomal transfer of PFF does not depend on bulk endocytosis
We then moved to other markers of endocytosis to explore the identity of PFF internalization. Due to the rapid transport of PFF to lysosomes, we also sought to determine whether phagocytosis may be responsible for uptake of PFF. RAW 264.7 macrophage cells were activated with LPS (1 μg/ml) 24 h prior to the experiment. Cells were given fluorescent latex beads (FluoSpheres) with PFF. Little colocalization was observed between FluoSpheres and PFF (Fig S4A). We then used dextrans of different molecular weights were administered to U2OS cells alongside or prior to the addition of PFF (Fig S4B). We then moved to a macropinocytic marker, dextran (Fig S4C). Dextran with different molecular weights were used, due to the relationship between the size of dextran and its endocytic pathway (Li et al., 2015). The colocalization was only 25% when 10k dextran was added simultaneously with PFF (Fig S4D). Higher molecular weight dextran samples showed lower colocalization when added simultaneously with PFF. Thus PFF do not utilize bulk entry but seem to be internalized through a unique form of endocytosis.
Endocytosis of PFF is clathrin-independent
The current consensus is that PFF enter cells via CME (Uemura et al., 2020). However, we are aware of no mechanism by which cargo that enters cells via CME can gain access to lysosomes in 2 min while bypassing early endosomes. To test if CME is involved in the endocytosis of PFF, we used an established genetic approach involving the knockdown of CHC with previously validated siRNA (Galvez et al., 2007; Kim et al., 2011). Immunoblot reveals effective CHC knockdown in U2OS cells (Fig. 4A). We then plated cells treated with control siRNA or CHC-targeting siRNA as a mosaic in the same well and incubated them with PFF. We observed no difference in internalization of PFF when comparing knockdown and control cells (Fig. 4B/C). Thus, it does not appear that CME is a major route for the internalization of PFF.
Internalization of PFF occurs through macropinocytosis
Holmes et al. (2013) demonstrated that tau fibrils utilize macropinocytosis for cellular entry and that fibril α-syn colocalizes with tau during uptake, suggesting a potential role for macropinocytosis in the internalization of fibril α-syn. Consistently, Zeineddine (2015) (Zeineddine et al., 2015) found that α-syn fibrils induce membrane ruffling, an early step in forming macropinosomes. Although macropinocytic cargo generally follow the endosomal pathway (Mayor and Pagano, 2007), we sought to test if macropinocytosis is involved in the internalization of PFF. We first examined if EIPA, an established inhibitor of macropinocytosis (Commisso et al., 2014; Koivusalo et al., 2010; Nakase et al., 2015) influences the internalization of PFF. EIPA disrupts the Na+/H+ exchanger, decreasing cytosolic pH and inhibiting the activation of Cdc42 and Rac1, required for macropinocytosis (Koivusalo et al., 2010). At a concentration of 20 µM, EIPA inhibits uptake of PFF in human astrocytes (Fig. 4D/E). Latrunculin B (LatB), which inhibits macropinocytosis by disrupting actin polymerization (Williams and Kay, 2018), had a similar block on uptake of PFF when used at 5 µM (Fig. 4D/E). In contrast, neither drug influenced the uptake of EGF (Fig. 4D/F), which at the concentration used, enters cells via CME (Sigismund et al., 2005). In addition to astrocytes, we confirmed our findings in dopaminergic NPCs. Latrunculin A (LatA), which like LatB is a potent actin polymerization inhibitor (Fujiwara et al., 2018), demonstrated a dose-dependent inhibition of PFF uptake in dopaminergic NPCs with uptake reduced by 81% compared to control at 2 μM (Fig. 4G).
Consistent with a role for macropinocytosis in endocytosis of PFF, PFF induce the formation of actin-rich membrane ruffles, precursors of macropinosomes (Condon et al., 2018) (Fig. S5A/B). More specifically, PFF induce recruitment of Rac1 to actin-rich membranes on the cell surface, a characteristic of macropinocytosis (Grimmer et al., 2002) (Fig. S5A). The ability of PFF to induce membrane ruffling is also readily observed by EM (Fig. S5C). As a control, we examined the influence of EGF, a documented inducer of membrane ruffles, even though it does not use macropinocytosis itself for internalization (Bryant et al., 2007). We also examined Tf, which is not known to induce membrane ruffling or macropinocytosis. As expected, EGF causes the recruitment of Rac1 to the surface, where it colocalizes with F-actin, whereas no Rac1 recruitment is seen upon the addition of Tf (Fig. S5D). Thus, PFF appear to stimulate membrane ruffles and use macropinocytosis to gain direct access to lysosomes.
Trafficking itinerary of PFF revealed by immunogold EM
To observe the trafficking itinerary of PFF directly we labeled the fibrils with gold (Fig. S1C) and followed their trafficking by EM. At 2-3 min following addition of PFF, astrocytes were fixed and processed for EM with uranyl acetate staining. Gold-labeled PFF appear under membrane ruffles and intracellular macropinosomes formed by ruffle closure (Fig. 5Ai/ii). The immunogold-labeled PFF remain in the lumen of the macropinosomes/endosomes that begin to demonstrate inward invagination of vesicles (Fig. 5Aiii/iv). At 3-5 min the PFF can be found in the lumen of intracellular membranes that gradually acquire electron density, indicating that they are likely lysosomes (Fig. 5C; Fig. S6A/B).
Remarkably, gold-labelled PFFs are often found in close association with regions of membrane curvature as they begin to bud inward into electron-lucent organelles approximately 300-400 nm in diameter (Fig. 6A). Thus, PFF may contribute to the formation of MVBs. In fact, PFF are readily detected on the surface of vesicles in MVBs (Fig. 6B). MVBs, positive for LAMP1, containing PFF are also detected at the light level using STED microscopy and fluorescently-labeled PFF (Fig. 6C). Live-cell studies were performed to examine the dynamics of these structures, which appear to undergo multiple rounds of fusion and membrane budding (Fig. 6D/E). Thus, PFF enter cells via macropinocytosis and appear rapidly in lysosomes and MVBs.
PFF are transferred to naïve cells using exosomes
The spread of PFF throughout the nervous system requires that the fibrils escape cells and be transferred to neighboring naïve cells. Exosomes provide a mechanism for cell-to-cell transfer of protein, lipids, and other cellular molecules, and the presence of PFF in MVBs suggests a potential mechanism for cellular release. To test this hypothesis, U2OS cells with stable expression of CD63-GFP were incubated with fluorescently-labelled PFF for 24 h and then washed with trypsin and buffer before re-plating with PFF-naïve cells with stable expression of LAMP1-RFP. From 12-24 h following the start of co-culture, PFF were observed to transfer from the CD63-positive cells to the LAMP1-positive cells (Fig. 7A/C). Moreover, we also observed transfer of CD63 (Fig. 7A/B), suggesting that transfer of PFF involves exosomes. Similar results are seen with transfer to naïve cells expressing CD9-RFP (Fig. S7A). Notably, the transfer of both CD63 and PFF is blocked by the addition of manumycin A, a compound that disrupts exosome release (Fig. 7A-C).
We next took CD63 donor cells and incubated them with gold-labeled PFF for 24 h before extensive trypsin/buffer wash. The cells were then re-plated in serum-free media and incubated for 36 h before the media was collected, spun at low speeds to remove any cells or cell debris, and the supernatant transferred to naïve cells. We were able to capture images of extracellular vesicles in PFF naïve cells using confocal microscopy (Fig. 7D). We also found that the internalization of these exosomes was blocked by using EIPA (Fig. 7E). Following the same protocol to prepare EM samples, with the exception of a 48 h incubation time in fresh media supplemented with serum following re-plating, we found exosomes decorated with PFF in contact with naïve cells, sometimes seemingly near sites of membrane ruffles and forming macropinosomes (Fig. 7F). In investigating the response of cells to PFF, CD63-GFP expressing cells were given PFF for 24 h and were then washed with trypsin and re-plated. These cells often revealed CD63-positive ruffling and at the tip of ruffles, we could often spot PFF puncta colocalizing with CD63 (Fig S7B). We then isolated exosomes from WT U2OS cells that were exposed to PFF for 24 h. Exosomal isolation was performed using a series of centrifugation steps (Chhoy et al., 2021). Exosomes isolated from PFF treated and control cells showed fibril-like structures on their surface while the control exosomes did not (Fig S7C). Additionally, a portion of the exosomes isolated from PFF treated cells were trypsinized to determine whether PFF remains exclusively on the exosomal surface. Immunoblotting of the exosomal samples not only proved the identity of the exosomes (CD63-positive, GM130-negative) but additionally showed that the trypsinized exosomes did not contain any alpha-synuclein, hence PFF only reside on their surface (Fig S7D). Lastly, exosomes were isolated from U2OS cells that were given gold PFF for 24 h, trypsin washed, and then incubated for 48 h in fresh media (Fig S7E). Low centrifugation speeds allowed us to isolate exosomes, thanks to the nano-gold PFF, and allowed us to visualize them using EM. Gold covered exosomes were isolated in large clusters; however, we were able to show that PFF remains exclusively on the surface of exosomes. In summary, PFF that have entered cells by macropinocytosis, are trafficked to MVBs and appear to be delivered on exosomes to neighbouring cells.
DISCUSSION
The mechanisms regarding the uptake of PFF remain incomplete (Bieri et al., 2018; Grozdanov and Danzer, 2018). Most studies investigating internalization of α-syn or PFF evaluate uptake hours following addition to cells (Desplats et al., 2009a; Hansen et al., 2011; Konno et al., 2012; Lee et al., 2008; Liu et al., 2007; Luk et al., 2016; Luk et al., 2009; Madeira et al., 2011; Rodriguez et al., 2018; Sung et al., 2001; Volpicelli-Daley et al., 2014; Volpicelli-Daley et al., 2011). While such uptake assays are valuable for genetic screening, they lack the temporal resolution to understand the pathways involved in endocytosis, an early event. We performed a detailed analysis of early events in endocytosis of PFF and were surprised to discover a unique uptake mechanism that allows the protein to reach lysosomes within 2 minutes, bypassing the early endosomal system.
Endocytosis of PFF has been thought to follow canonical CME pathways, entering cells via clathrin-coated pits and vesicles followed by trafficking through endosomes to lysosomes (Konno et al., 2012; Lee et al., 2008). However, we found no evidence for trafficking of PFF through early or recycling endosomes, inconsistent with a CME pathway. Moreover, using a previously established pool of CHC siRNA (Bayati et al., 2021; Galvez et al., 2007; Kim et al., 2011), we attained ∼95% knockdown efficiency, with no influence on PFF endocytosis. Thus, our data does not support a role for CME in uptake of PFF.
PFF induce the formation of actin- and Rac1-rich membrane ruffles, precursors of macropinocytic vesicles (Cox et al., 1997; Grimmer et al., 2002). Further, inhibitors of macropinocytosis (Commisso et al., 2014; Koivusalo et al., 2010; Zwartkruis and Burgering, 2013) including EIPA, LatA and LatB (Erami et al., 2017; Furstner et al., 2007; Morton et al., 2000; Wakatsuki et al., 2001) significantly inhibited uptake of PFF while not affecting EGF uptake. Our findings point to endocytosis of PFF through a form of macropinocytosis, where macropinosomes either mature into MVBs and lysosomes or fuse with pre-existing LAMP1-positive compartments. Recent papers show promising results regarding the role of actin in PFF internalization (Underwood et al., 2020; Zhang et al., 2020).
Aspects of our findings corroborate previous studies. First, although slower, previous research demonstrates the direct fusion of macropinosomes with lysosomes (Yoshida et al., 2018). Second, the findings of previous studies on dynamin inhibition and reduction in uptake of PFF can be due to dynamin’s role in some forms of macropinocytosis and its involvement in actin remodelling (Gu et al., 2010; Krueger et al., 2003; Mulherkar et al., 2011). In fact, at high concentrations, we observed that dynasore blocks internalization of PFF (data not shown).
Finally, amilorides, such as EIPA, which block uptake of PFF have been shown to have neuroprotective effects in PD (Arias et al., 2008).
Although previous studies attempted to examine the localization of exogenous α-syn using immuno-EM (Volpicelli-Daley et al., 2011), we conjugated PFF directly with gold. Many antibodies to α-syn cannot distinguish between different conformations of the protein (Kumar et al., 2020), making the direct conjugation of gold to PFF a more specific method. Consistent with previous literature (Vargas et al., 2017), we found that PFF were almost always in close association with membranes, whether at the cell surface, during internalization, or while in MVBs and lysosomes.
At early time points after addition to cells, we observed gold-PFF on inward invaginating vesicles in MVBs. At longer time points, gold-PFF were seen to be transported on exosomes, which confocal microscopy, demonstrated were CD63 positive. Two important discoveries were made as a result of this: first, PFF transmission is at least partly due to the exosomal transport of PFF; secondly, PFF reside on the surface of exosomes, rather than being contained in the lumen of exosomal vesicles. Further, exosomes isolated from cells exposed to PFF contained PFF on their outer surface. Biochemically, we were able to show that PFF can be removed from exosomes through trypsinization; this would only be possible if PFF resided on the surface of exosomes. We then used Manumycin A, a drug that blocks the release of CD63-positive exosomes (Datta et al., 2017), and found that Manumycin A disrupts PFF transmission from cell-to-cell. We suspect that since PFF resides on the outside of exosomes, that drugs like EIPA, LatA, and LatB that blocked its initial uptake, can be used to block PFF spread from cell-to-cell. Our hypothesis was confirmed with our experiment using EIPA, showing the inhibition of exosomal uptake by PFF-naïve cells.
Although all evidence presented indicates that PFF use macropinocytosis to gain access into cells, we are also cognizant that the properties of α-syn itself could drive the membrane curvature and protrusions. As previously observed, PFF associate and may even drive membrane curvature (Vargas et al., 2017; Westphal and Chandra, 2013). This could mean that PFF drives its own internalization into the cell by causing membrane curvature, and then, once trafficked to MVBs/lysosomes, it drives membrane invaginations, resulting in exosomal formation and eventually its release. Once released, PFF then goes on to disrupt other cells, all the while remaining on the outside of exosomes, allowing it to drive membrane curvature that would drive its internalization in neighboring cells. Although PFF is only an oligomer, it certainly has many prion-like characteristics. Above all else, PFF may be taking advantage of the α-syn’s interaction with membranes to drive its own endocytosis, exocytosis, and drive the production of more aggregated α-syn. It is this evolutionary drive for PFF to create more fibrillated forms of α-syn, that truly makes it more than just an oligomer, and more like a prion.
In conclusion, our data demonstrate that PFF enter cells via macropinocytosis, with seemingly direct transfer to MVB and lysosomes, bypassing early endosomal pathways. It remains unclear if this represents fusion of macropinosomes with MVB and lysosomes or maturation of macropinosomes into these structures. A portion of the PFF that enter the cells are subsequently secreted on exosomes, providing a mechanism for cell-to-cell transport, and yet our data does not explain how fibrillar αSyn interacts with αSyn in the cytosol to allow prion-like propagation. Regardless, inhibition of macropinocytic pathways may prove useful in limiting the progression of PD and other synucleinopathies.
AUTHOR CONTRIBUTIONS
A.B. planned and conducted the experiments and wrote the manuscript with P.S.M. E.B. performed lysosomal immunoprecipitation experiment along with analyzing protein expression through western blotting. C.H. provided NPCs and differentiated iPSCs into cortical and dopaminergic neurons. W.L. produced and characterized PFFs. W.E.R. conducted 24 h PFF uptake experiment with dopaminergic NPCs using LatA. C.Z. provided us with additional NPCs and differentiated iPSCs. R.S. measured and analyzed PFF length from multiple batches of PFF using dynamic light scattering. E.D.C.P. characterized and measured PFF size via EM. B.V. suggested experiments investigating macropinocytosis. H.M.M., E.A.F. and T.M.D. edited the manuscript and helped provide overall direction of the project. P.S.M. funded and supervised the project, aided A.B. in planning the experiments, and wrote the manuscript with A.B.
DECLARATION OF INTERESTS
The authors declare no competing interests.
STAR METHODS
KEY RESOURCE TABLE
RESOURCE AVAILABILITY
Lead Contact
Further information and requests for resources and reagents should be directed to the Lead Contact, Dr. Peter Scott McPherson (peter.mcpherson{at}mcgill.ca) and will be fulfilled.
Materials Availability
This study did not generate new unique reagents.
Data and Code Availability
All the raw data, along with statistical calculations used in this paper have been submitted in a spreadsheet file along with the manuscript to Cell Press. All data reported in this paper will be shared by the lead contact upon request.
This paper does not report original code.
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.
EXPERIMENTAL MODELS AND SUBJECT DETAILS
Cell lines
HeLa, U2OS, U87, U-343, RAW 264.7 were obtained from American Type Culture Collection (Cat# CRM-CCL-2, HTB-96, HTB-14, discontinued, TIB-71, respectively). U-251 cells were obtained from Sigma (Cat# 09063001). Human fetal and adult astrocytes were obtained from Cell Applications (Cat# 882AK-05f and 882AK-05a, respectively). For studies with iPSCs, we used the line AIW002-2 obtained from the Neuro’s C-BIG Biorepository. This line was reprogrammed from peripheral blood mononuclear cells of a healthy donor with the Cytotune reprogramming kit (ThermoFisher, Cat# A34546). The process of reprogramming and quality control profiling for this iPSC was outlined in a previous study (Chen et al., 2021). The use of iPSCs in this project is approved by the McGill University Health Centre Research Ethics Board (DURCAN_IPSC / 2019-5374).
All cells were cultured in DMEM high-glucose (GE Healthcare, Cat# SH30081.01) containing 10% bovine calf serum (GE Healthcare, Cat# SH30072.03), 2 mM L-glutamate (Wisent, Cat# 609065, 100 IU penicillin and 100 μg/ml streptomycin (Wisent cat# 450201). Cell lines were routinely checked for mycoplasma contamination using the mycoplasma detection kit (Biotool cat# B39038).
Production, Characterization, and Nano-Gold labelling of PFF
Production and characterization of recombinant α-syn monomers and PFF have been described previously (Del Cid Pellitero et al., 2019; Maneca et al., 2019). Both electron microscopy and dynamic light scattering were used for the characterization of α-syn monomers and PFF (Fig. S1A-B). Previously characterized PFF, was then conjugated with 5 nm gold beads (Cytodiagnostics, Cat# CGN5K-5-2), immediately before experimental use. Cytodianostic’s conjugation protocol was used to conjugate gold onto PFF. Following conjugation, some PFF was collected for characterization on carbon-covered grids (Electron Microscopy Sciences, Cat# FCF400CU50) (Fig. S1C).
IPSC Culturing
AIW002-2 hiPSC cultures were maintained as feeder-free cultures following the protocol described previously (Chen et al., 2021). AIW002-2 hiPSCs were plated onto Matrigel (Corning, Cat# 354277)-coated plates containing mTeSR1 medium (Stemcell Technologies, Cat# 85857). The culture medium was changed daily until cells reached ∼80% or required confluency (usually 5-7 days after plating). The cells were then passaged, frozen, or differentiated. A previously described protocol was used to generate ventral midbrain dopaminergic neural progenitor cells (Jefri et al., 2020). Dopaminergic neural progenitor cells were dissociated with StemPro Accutase Cell Dissociation Reagent (ThermoFisher, Cat# A1110501) into single-cell suspensions. 50,000 cells were plated onto coated coverslips in 24-well plates with neural progenitor plating medium (DMEM/F12 supplemented with N2, B27 supplement; ThermoFisher, Cat# A4192001, 17502001, 17504044). To further differentiate into dopaminergic neurons, neural progenitor medium was switched to dopaminergic neural differentiation medium (Brainphys Neuronal medium, STEMCELL Technologies, Cat# 05790) supplemented with N2A Supplement A (STEMCELL Technologies; Cat# 07152), Neurocult SM1 Neuronal Supplement (STEMCELL Technologies; Cat# 05711), BDNF (20Lng/mL; Peprotech, Cat# 450-02), GDNF (20Lng/mL; Peprotech, Cat# 450-10), Compound E (0.1 μM; STEMCELL Technologies, Cat# 73954), db-cAMP (0.5 mM; Carbosynth, Cat# ND07996), Ascorbic acid (200 μM; Sigma Aldrich, Cat# A5960) and laminin (1 μg/mL, Sigma Aldrich, Cat# L2020).
Plasmids and lentivirus
EBFP2-Lysosomes-20, tdTurboRFP-Lysosomes-20 were gifts from Michael Davidson (Addgene plasmid# 55246 and 58061). pLJM1-Tmem192-mRFP-3xHA was a gift from Roberto Zoncu (Addgene, plasmid# 134631). pLJC5-Tmem192-3xHA was a gift from David Sabatini (Addgene plasmid # 102930). CD63-pEGFP C2 was a gift from Paul Luzio (Addgene, Cat# 62964). LAMP1-RFP lentivirus was obtained from Applied Biological Materials Inc. (Cat# LVP719).
METHOD DETAILS
Trypan Blue Exclusion Assay
HeLa (ATCC, Cat# CRM-CCL-2) cells were grown on poly-L-lysine coated 35 mm glass-bottom dishes (MatTek, Cat# P35G-1.5-14-C-GRD) for 48 h to 50% confluency. When ready for imaging, cells were stained with Hoescht (Invitrogen, Cat# H3570) for 30 min and dishes were placed in the pre-heated live imaging chamber of the Zeiss LSM-880 confocal microscope.
Imaging was commenced before the addition of PFF tagged with Alexa Fluor 488 (PFF488) at a speed of 1 frame/sec. PFF was added during live imaging.
PFF Live Internalization Assay
U2OS (ATCC, HTB-96) cells were grown on Poly-L-Lysine (Sigma Aldrich, Cat# A-005-M) coated 35 mm glass-bottom dish (MatTek, Cat# P35G-1.5-14-C-GRD) for 72 h to 70% confluency. Before imaging, cells were stained with Lysosomal Cytopainter (Abcam, Cat# ab176827) for 30min. Dishes were placed in the pre-heated live imaging chamber of the Zeiss LSM-880 confocal microscope. Imaging was commenced before the addition of PFF at a speed of 1 frame/sec using the Airy imaging mode for higher resolution (∼1.3x resolution of conventional confocal microscopy). PFF488 was added during imaging.
Live MVB Assay
These live experiments were carried out the same way as above. For the hollow lysosome morphology, the “Find Edges” processing was used on ImageJ (NIH, https://imagej.net/software/fiji/).
Lysosomal Staining in Fixed Samples
Lysosomal staining was achieved in one of the following ways: transfection with plasmids stated above with the aid of Lipofectamine 3000 (ThermoFisher, Invitrogen, Cat# L3000015), fixable Lysotracker (ThermoFisher, Invitrogen, Cat# L7528) for staining of lysosomes in neurons and neural progenitor cells, or with the use of LAMP1 antibody (Cell Signaling, Cat# 9091S. Transfection was done 24 h prior to experimentation. Staining with Lysotracker was done 30 min before experimentation. LAMP1 antibody staining was done following fixation and permeabilization.
PFF Endocytosis Assays
Cells mounted on coverslips were plated in 24 well plates at 37°C with 200 µl of media in each well. Immediately prior to experimentation, cells were taken out of incubators and placed in the cell culture hood. PFF aliquots were then removed from dry ice, diluted with serum-free media, and added to each well. Cells were then incubated at 37°C for 0, 2, 10, 30, or longer. Cells were then placed on ice and washed with trypsin (Wisent, Cat# 325-052-EL) for 90 sec (to remove extracellular PFF). Following 2 washes with PBS (Wisent, Cat# 311-010-CL), cells were fixed. In experiments where trypsin was not used, cells were washed three times with PBS and fixed. In experiments using iPSC neurons, trypsinization was avoided and only PBS was used to wash cells. Same protocol was followed for samples prepared for EM, except for plating cells on Nunc 8-well dishes (Lab-Tek, Nunc, Thermo Scientific, Cat# 177445) and fixation with Glutaraldehyde and not PFA.
PFF Long-term Endocytosis Assays
Cell plating and culturing was done as described above. Cells were then incubated at 37°C for 24 h. Following 24 h, cells were trypsin washed three times, pelleted, resuspended and re-plated onto fresh coverslips. For iPSC neurons, trypsinization (Wisent, Cat# 325-052-EL) and re-plating was avoided and only the media was changed at 24 h. Cells were then incubated for varying amount time afterwards. Prior to imaging, sample fixation and mounting were conducted as above.
For quantifying the PFF fluorescence over 14 d, the same protocol as above was employed; however, in order to control for cell growth, human adult astrocytes (Cell Applications, Cat# 882AK-05a) at passage 7 were used as these cells have a very slow doubling time. This allowed us to control for growth without needing to inhibit cell replication through treatment with chemicals that could have a negative affect on PFF uptake, fluorescence, or retention.
Macropinocytic Inhibitors
Ethylisopropyl amiloride (EIPA; R&D Systems, Tocris, Cat# 3378), Latrunculin A (LatA; Cayman Chemical, Cat# 10010630), and Latrunculin B (LatB, respectively; abcam, Cat# ab144920, ab144291) were the macropinocytic inhibitors used in this study. Cells normally incubated in media with serum were washed 3 times with serum-free media to remove any remaining serum. Serum-free media was added with a final concentration of either 20 µM of EIPA, 0-2 µM of LatA, 5 µM of LatB. Cells were then incubated at 37°C for 30 min before experimentation for EIPA and LatB, and 1 h for LatA. Due to the autofluorescence of EIPA, DRAQ7 (abcam, Cat# ab109202) was used to stain cell nuclei.
LatA Inhibition in Dopaminergic NPCs
Dopaminergic NPCs were seeded in polyornithin/ laminin coated 384 well plates (Corning, Cat# 353962) at 4000cells/ well. After 24 h Latrunculin A (LatA; Cayman Chemical, Cat# 10010630) was added. After 1 h, Alexa488-PFF was added. After another 24 h incubation period, the wells were washed twice with PBS (Wisent, Cat# 311-010-CL) and the cells fixed with 2% FA/PBS.
Cells were counterstained with Hoechst 33342 (ThermoFisher, Cat# H3570) and imaged on a high content imager (CellInsight CX7, ThermoFisher Scientific). The amount of intracellular Alexa488-PFF was measured as the total intensity of Alexa488 fluorescent stain in the perinuclear area. Nuclei count and nuclear area (px2) were also obtained as indicators for cell toxicity. All data was normalized against DMSO (vehicle) only controls. Data represents the mean and standard deviation of 3 independent experiments.
Lysosomal Immunoprecipitation
U2OS cells stably expressing 3xHA-TMEM192-RFP were incubated with PFF for varying lengths of time, washed with trypsin (Wisent, Cat# 325-052-EL) three times to remove extracellular PFF, and then trypsinized and pelleted. Lysosomes were immunoprecipitated based on the protocol described in Abu-Remaileh et al. (2018). Briefly, cells were washed twice with KPBS and centrifuged at 4°C for 2 minutes at 1000 x g. Pelleted cells were resuspended in lysis buffer (20 mM HEPES pH 7.4, 100 mM NaCl, and protease and phosphatase inhibitor cocktail) and manually homogenized with 25 up and down strokes. Homogenates were centrifuged for 2 min at 4°C at 1000 x g and 10% of the total volume for each IP was reserved for starting material (SM) fractions. Homogenates were incubated with gentle rocking at 4°C for 3 minutes with 100 μL anti-HA magnetic beads (ThermoFisher, Thermo Scientific, Cat# 88836) pre-washed with KPBS. After collecting the total IP volume for the unbound materials (UM), immunoprecipitates were washed three times with KPBS using a DynaMag-2 Magnet (ThermoFisher, Invitrogen, Cat# 12321D) before resuspension in 1X SDS-PAGE sample buffer. Protein aliquots were then analyzed by SDS-PAGE and Immunoblot.
For samples to be processed via confocal microscopy, 20 μL pre-washed anti-HA beads were used in the IP and were resuspended in using PFA to resuspend and fix them. The sample was then pelleted and resuspended with PBS (Wisent, Cat# 311-010-CL) and mounted onto slides.
PFF Intercellular (with Contact) Transfer Assay
U2OS cells were transfected with CD63-EGFP or LAMP1-tdTurboRFP using Lipofectamine 3000 transfection reagent and incubated for 24 h. Cell media was then changed, and cell selection process was initiated through the addition of Neomycin (Gibco cat# LS21810031). U2OS cells stably expressing CD63-GFP were exposed to PFF or PBS (vehicle control) for 24 h. CD63-GFP (donor cells) were then trypsin washed three times, pelleted, trypsinized again, pelleted and PBS washed, prior to being co-plated with acceptor cells (PFF naïve cells stably expressing LAMP1-tdTurboRFP). Donor and Acceptor cells were then incubated for 12 or 24 h. Half of the 24 h sample were given Manumycin A (MA; Cayman Chemical, Cat# 10010497) at 1.2 µM while the other half was given DMSO (vehicle control) at the time of plating. Cells were then fixed, mounted and imaged.
PFF Intercellular (without Contact) Transfer Assay
U2OS cells were stably expressing with CD63-EGFP, the cells were exposed to PFF, for 24 h then trypsin washed three times and re-plated prior to the addition of serum-free media. The media was collected following 36 h of incubation at 37°C and given to naïve U2OS cells. Some of the cells given this media were exposed to EIPA while others were given DMSO as a vehicle control. For fluorescent microscopy samples, the naïve cells were plated on coverslips, while for samples to be examined with EM, naïve cells were plated on Nunc 8-well plates (Lab-Tek, Nunc, Thermo Scientific, Cat# 177445).
Fixation and Antibody Staining Following PFF Uptake
Fixation was done with 2% freshly made paraformaldehyde (PFA; Thermo Fisher Scientific, A1131322) for 10-15 min on ice. In experiments where antibody staining was done, cells were then blocked and permeabilized for 30 min using 0.05% Triton X-100 (Sigma, Cat# X100-1L) in Phosphate-buffered saline (PBS; Wisent, Cat# 311-010-CL) along with 5% BSA (Wisent, Cat# 800-095). Coverslips were then transferred into a wet chamber and incubated with 1:500 dilution of the antibody in 0.01% Triton X-100 and 5% BSA. Cells were incubated with the diluted antibody for 2 h at room temperature. Coverslips were then gently washed 3 times with PBS, and 1:500 dilution of secondary antibody was added in 0.01% Triton X-100 and 5% BSA. Cells were then washed 2 more times with PBS and stained with DAPI (ThermoFisher, Invitrogen, Cat# D1306) for 10 min at 1 µg/ml concentration. Coverslips were then mounted on a glass slide using Fluorescence Mounting Medium (Dako, Agilent, Cat#S3023). All fixed samples were then imaged using a Leica TCS SP8 confocal microscope. STED samples were imaged using Abberior STED super-resolution nanoscope.
We initially found that PFF staining was severely affected following permeabilization limiting our ability to use antibodies in our experiments. We found that in cells with higher uptake of PFF (e.g., astrocytes), we could permeabilize samples and still retain some PFF fluorescence. We also found that instead of fixation and permeabilization with methanol or PFA and high concentrations of Triton X-100 (i.e., 0.1-0.5% Triton volume/volume), that a softer fixation with 2% PFA and low concentration of Triton X-100 for permeabilization (i.e., 0.05% Triton) allowed for the preservation of PFF fluorescence. Following our understanding of how to mediate the deleterious effects of permeabilization on PFF fluorescence in fixed samples, we could then employ antibodies in our experiments to stain for endosomal markers.
CHC KD
U2OS cells at 60% confluency were transfected with siRNA retrieved from Dharmacon (SMARTpool, ONTARGETplus, see KRT) or control siRNA (Dharmacon; ON-TARGETplus CONTROL) using Lipofectamine 3000 (ThermoFisher, Invitrogen, Cat# L3000015). At 24 h following transfection, cells were passaged, with some cells retrieved for imaging experiments and mounted onto coverslips. At 48 h, cells were collected in HEPES lysis buffer (20 mM HEPES, 150 mM sodium chloride, 1% Triton X-100, pH 7.4) accompanied with protease inhibitors. Cells in lysis buffer were then placed at 4°C and gently rocked for 30 min. Lysates were then spun at 238,700 x g for 15 min at 4°C. Lysates were run on 5-16% gradient polyacrylamide gel and transferred onto nitrocellulose membranes. Subsequently, proteins were visualized using Ponceau staining. Blots were then blocked with 5% milk for 1 h. Antibodies were then incubated overnight at 4% in bovine serum albumin/TBS/0.1 Tween 20. The secondary antibody was incubated at 1:2500 dilution in 5% milk/TBS/0.1 Tween 20 for 1 h at room temperature. Concurrently, cells mounted onto coverslips underwent the endocytosis assay with both knockdown and control cells mounted onto the same coverslips.
Fluorosphere and Dextran Uptake Assays
RAW 264.7 cells (ATCC, Cat# TIB-71) were plated onto coverslips treated with Poly-L-Lysine (Sigma, Cat# A-005-M). FluoSpheres (ThermoFisher, Invitrogen, Cat# F13082), latex beads of 1 µm diameter, and orange fluorescence were incubated with and without PFF (before and simultaneously). Cells were then washed with trypsin (Wisent, Cat# 325-052-EL), washed twice with PFF, and fixed. Cells were then counterstained with DAPI prior to mounting using Fluorescence Mounting Medium (Dako, Agilent, Cat#S3023).
U2OS cells were used in the dextran and PFF co-uptake experiment. U2OS cells were plated and prepared as described above. Three different dextran (10,000 MW, 70,000 MW, 2,000,000 MW; ThermoFisher, Cat# D1817, D1818, D7139) were separately added to cells prior or with PFF addition.
Membrane Ruffling Assay
Cellular response to PFF addition was done using human fetal astrocytes (Cell Applications, 882AK-05f). Cells were treated with PFF, and their recruitment of Rac1 (Emd Millipore Corporation, Cat# 16319) to the cell surface and colocalization with F-actin (phalloidin; Abcam, Cat# A22287) was analyzed. For negative and positive control, Transferrin (Tf; ThermoFisher, Invitrogen, Cat# T13342) Epidermal growth factor (EGF; ThermoFisher, Invitrogen, Cat# E13345) were added to cells, respectively. A null condition was also in place, were cells were administered PBS (a vehicle control for PFF).
Exosomal Isolation
For exosomal isolation, a previously published protocol was followed (Chhoy et al., 2021). Following isolation, samples were prepared for immunoblotting and EM. For immunoblotting, a portion of exosomes retrieved from PFF treated cells were treated with trypsin for 5 min. The exosomes were then re-pelleted at 10,000 g’s, and the supernatant removed. These trypsinized exosomes were run alongside exosomes collected from PBS and PFF treated cells, to compare PFF content. For isolation of exosomes collected from cells exposed to PFF-gold, a single centrifugation step at 10,000 g’s was used to pellet exosomes.
TEM
Human astrocytes and U2OS cells were plated onto 8 well chamber slides (Lab-Tek, Nunc, Thermo Scientific, Cat# 177445) and were administered PFF conjugated with gold. Cells were then fixed with glutaraldehyde 2.5% in 0.1M sodium cacodylate buffer (Electron Microscopy, Cat# 1653715), post-fixed with 1% OsO4 and 1.5% potassium ferrocyanide in sodium cacodylate buffer. Cells were then en bloc stained with 4% uranyl acetate. Post-embedding, some grids were stained with uranyl acetate for enhanced membrane staining. Samples were viewed with a Tecnai Spirit 120 kV electron microscope and captured using a Gatan Ultrascan 4000 camera.
Graphical Abstract
Created with BioRender.com
QUANTIFICATION AND STATISTICAL ANALYSIS
Colocalization Rate Calculation
Colocalization was quantified using the Leica LAS X software. A 50% background threshold was set to eliminate all background fluorescence and fluorescent bleed-through across channels. Colocalization rate was used for graphs and statistical calculations. Colocalization was calculated by measuring the colocalization area between two channels (in µm2) and dividing it by the foreground area (foreground area = image area divided by total area). This allows us to look at what portion of the image is taken up by colocalization, a great indicator of the degree of colocalization between two channels, while considering the area that each channel occupies.
This is based on MOC, mander’s overlapp coefficient.
Quantification and Statistics
For all quantifications, including uptake and colocalization, the Leica LAS X and ImageJ (imagej.net) software were used. Readout from the LAS X software included Pearson Correlation Coefficient and colocalization rate expressed in percentage. In each experiment, images were selected from a large field imaged at low quality using the Leica SP8 spiral scan. From the large field, regions were selected that contained a minimum of 4 cells. Graphs were then prepared using GraphPad Prism 9 software. For statistical comparisons, Student’s t-test (unpaired) and one-way ANOVA were used. When significance was detected under ANOVA, multiple comparisons Tukey’s test was conducted to find means that are significantly different across conditions. All data is shown as mean ± SD when quantifying uptake and mean ± minimum and maximum values for colocalization. For statistical significance, p < 0.05 was used.
ACKNOWLEDGEMENTS
We acknowledge the Neuro Microscopy Imaging Centre and Advanced BioImaging Facility and the Facility for Electron Microscopy Research at McGill University. We thank Dr. Michael Davidson and Dr. Paul Luzio for the LAMP1 and CD63 plasmids, respectively. We also thank Drs. Sabatini and Zoncu for the HA-TMEM192 plasmids. AB is supported by Fonds de recherche du Québec doctoral award and a studentship from the Parkinson Society of Canada. This work was supported by a grant from the Canada First Research Excellence Fund, Healthy Brain, Healthy Lives, McGill University, awarded to PSM. PSM is a Distinguished James McGill Professor and Fellow of the Royal Society of Canada.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.
- 36.↵
- 37.↵
- 38.↵
- 39.
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.
- 64.
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.
- 77.
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.
- 85.↵
- 86.↵
- 87.
- 88.↵
- 89.
- 90.↵
- 91.
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.
- 112.↵